Stanford Researchers Launch Biomni, a General-Purpose Biomedical AI Agent

Stanford University researchers have released Biomni, a general-purpose AI agent designed to autonomously execute complex biomedical research tasks. The system, described as a "co-scientist," can interpret multimodal datasets, generate experimental protocols, and carry out computational analyses that would typically require a team of scientists days or weeks to complete. Biomni combines a large language model with a tool-use framework that grants it access to biological databases, analysis pipelines, and simulation environments. In case studies, the agent designed gene-editing experiments, analyzed single-cell sequencing data, and proposed hypotheses for disease mechanism studies without human intervention beyond the initial research question. The researchers frame Biomni as a step toward AI systems that can participate in the full cycle of scientific discovery rather than assisting with isolated tasks. The code and documentation are available publicly, and the team intends to extend the agent's capabilities to wet-lab automation interfaces in future releases.